Transcribed notes from the trip: Livermore visit: Don Muellonhof Ucompany RISC Module Systems: Developed specs and gave it to them Needed minor mod of existing design/(board?) Bob Peterson actually built one Task: Parallel I/O for actuator control, std serial ports interface Mass was an issue: Originally needed 40 PALs; Replaced with 4 mass gate arrays LSI logic (big), Orbit Semiconductor, Zylog, NEC, IDT All have small feature sizes The board is all CMOS ~ 10 W (mainly CPU) CPU 66 MHz, Bus 33 MHz < 1 year time, the unit worked the first time (off the bat) Mem: Interleaved DRAM DRAM: Feature size is smaller, but charge is being maintained Tests show DRAM upsets ~ SRAM upsets Cannot use for long periods as there is no heat sinking on the CPU Regular chips -55 < case temp < 85 C Mil -55 < case temp < 135 C Placement near motors, may (will) need shielding Watchdog timer: In every controller Distributed Vs Single: Distribute it Better isolation if power is distributed Low power, 8 bit controllers: 8051 -> can buy a complete set with memory and serial interface; Power: ~2W Microcontrollers can be used to simulate the system IDT: 4600:64 bits -> 100 MIPS -> look into 3.3 V 3081: 33 MHz, 27-28 MIPS, 10-11 MFLOPS, 5 W - scalable PowerPC can be good Way to find controller makers is to go to CHIP maker and ask for primary customers Cost: $500k, including 6 (PAL) + 6 (G array) prototypes Of tyhis 150 k : To run a mass gate array on a line Daryl Pines: Gimbal: 50 uRad accy, steady state in lab platform Bandwidth = 10 Hz for small angles mass ~ 4 kg Inertia: 250g - 500g mirror Jitter stability: 10 uRad over 100 msec PID on each axis Designed for MSTI Disturbances: prob. vibration: 42 g rms, 0 - 200 Hz -(filter fus-airplane) - 1 Hz - 120 Hz Used Mil Standards & followed them Internationa Space Directory is recommended Companies: Dow Corning, 3M - (industrial, space rated, comer.) Honeywell IMU susceptible to: 10 g RMS, 20 Hz - 2000 Hz 12-14 mo for next Gimbal